In Praise of the Ugly Spreadsheet
The ugly spreadsheet does not look objective. That can be its most honest feature.
You can tell when a spreadsheet has lived a life
The column widths disagree. One total is bold, another is double-underlined and a third has been given a yellow background whose meaning nobody can quite remember. There is a tab called Assumptions, a tab called Old assumptions and, inevitably, one called Copy of assumptions.
The person sending it usually begins with an apology.
Sorry, it is a bit ugly.
Oddly, that can make me trust the file more than a polished dashboard full of carefully rounded cards. Not because mess is proof of accuracy. It plainly is not. I trust it because I can see that somebody made it, changed it, argued with it and left enough of the working behind for me to do the same.
The ugly spreadsheet does not look objective. That can be its most honest feature.
A grid that became almost anything
Excel turned 40 in 2025. In Microsoft’s anniversary account, the familiar grid appears as far more than an accounting tool: people have used it for meal planning, bills, games, quilt patterns, animations and even football-stadium displays.
That range explains something about the spreadsheet’s stubborn survival. It begins without much opinion about what the work should become. A blank grid can be a calculator, a list, a tiny database, a schedule, a model or a place to work out what the actual question is.
Most specialised software begins after the question has been settled. The spreadsheet is often where the settling happens.
This is why its ugliness can contain useful information. The untidy part may show where reality refused to fit the first model. A note beside a number may explain the exception that mattered. An unusually wide column may exist because somebody discovered, after three meetings, that the short description was misleading. Those marks are part of the record.
A spreadsheet shows its seams
Imagine a dashboard with one large number: 14.3%.
It is centred, coloured green and labelled conversion. It looks finished. But the important questions are elsewhere. Which customers were included? Were refunds removed? What period does it cover? Did somebody exclude a week when the tracking failed?
In a spreadsheet, the answer might be inelegant but reachable. You can click the cell, follow the formula, inspect the input tab and find a note saying that three days were removed because the source file was incomplete. Perhaps the note is written in red Comic Sans. It is still explaining more than the dashboard’s immaculate decimal place.
The dashboard has seams too. They are simply farther away: in a query, a transformation, a permissioned analytics tool or a ticket raised six months ago. With a spreadsheet, the working is often still within pointing distance.
This does not mean every cell needs to be exposed to every reader. It means the route from answer back to assumption should exist. A working spreadsheet often provides that route accidentally, because the presentation and the machinery occupy the same surface.
The useful mess has a shape
There is a particular kind of spreadsheet I like. It is plain but explainable. Inputs have names. Important adjustments have notes. Dates and versions are visible. Formulas can be followed without first locating the one person who understands the file.
It may still contain an over-bright warning cell or a tab that should have been deleted last Tuesday. That is different from a workbook held together by broken links, hidden columns and hope.
What matters is whether the file is legible or mysterious.
A highly designed sheet can be legible. A rough one can conceal a catastrophe. The point of praising the ugly spreadsheet is to notice what often gets removed during the journey towards a more professional-looking result: qualifications, alternatives, intermediate calculations and traces of disagreement.
Ugly is not the same as careless
Spreadsheets deserve their bad reputation too. They are easy to copy, hard to govern and capable of becoming business-critical long after the original author has forgotten why a formula works.
The UK Government’s guidance on creating and sharing spreadsheets makes the problem almost charmingly clear: even a file created for personal use may end up serving a wider audience. It recommends versioning, descriptive names, documentation for embedded code and checks for hidden or sensitive data. These are not cosmetic improvements. They help somebody else understand what they have inherited.
A usable spreadsheet can show its mess without requiring the next person to become an archaeologist. There is no virtue in an undocumented macro or a total that changes when a hidden row is unhidden.
But neither should we confuse the removal of visible mess with the removal of uncertainty. Sometimes the rough edges are the only warning that a number involved judgement somewhere along the way.
I am less impressed by a spreadsheet that looks like software than by one that behaves like a good working note. I want to see where the number came from, what changed and which assumption would alter the answer.
So when somebody sends me a file and apologises for the colours, I tend not to worry about the colours first.
I click the cell.
Sources
- Microsoft: Formulas, fandom and an AI-fuelled future — Microsoft Excel celebrates 40 years, published 30 September 2025.
- UK Government: Creating and sharing spreadsheets, published 3 June 2021 and last updated 5 October 2023.